This species, a common malodorous pantropic weed, merits attention here in light of Bentham’s13 remarks concerning its close affinity to C. sophera L. In Guiana any problem of confusion between the two species is as yet academic, for the latter has not been reported from the region. However, in view of its presumed origin in the paleotropics and known successful establishment in the vicinity of Caracas (Pittier 9712, 11563) and in Trinidad (Schack s.n.; Broadway 8814), the occurrence of C. sophera in Guiana may be expected.
On the basis of personal experience in the field with C. occidentalis and in cultivating both species under uniform garden conditions in Georgetown, British Guiana, for two years and in Austin, Texas, for three years, from seed obtained from a variety of tropical and subtropical sources in both the Old and New World, the following observations may be made, with special reference to Bentham’s discussion:
In general, C. occidentalis is remarkably uniform throughout its very extensive range. The scope of variation in the species given in Table 2, based on observations made of many plants, is frequently as valid for a single plant. Further, it was clearly noted that, irrespective of the source of the seed, the plants grown in Texas were strongly monocarpic, tending to diminish in vigor after 8 to 12 inflorescences on a given branch had matured fruit. Plants grown from seed sown in April showed marked reduction in growth by mid-September and had died long before the onset of cold weather. In British Guiana the decline in vigor occurred at the beginning of the long dry season (usually August) and was less abrupt. Occasional plants survived and, with the resumption of abundant soil moisture in late November, rejuvenated from the base. Corolla wilt, a common phenomenon in Cassia Sections Oncolobium, Prososperma, Absus, and Chamae-crista, occurred by 11 am on bright days in Texas. No relevant notes were made in British Guiana.
By contrast, C. sophera is highly variable in a number of characters, these having little relation to geography in the African, Asiatic, and Australian material seen, but more clearly localized in exsiccatae from tropical America. This
latter circumstance may be interpreted as having resulted from the introduction of elements of Old World populations which, now out of contact with the gene pool in the original areas, tend to stabilize. For example, the cited collections from Venezuela show greater affinity to certain Australian collections (Irwin 2832: cultivated in Austin, Texas, from seed sent from Adelaide) than to those of Trinidad.
Under garden conditions in Austin, Texas, plants of C. sophera all tended to have a more divergent, looser pattern of branching, to grow more slowly, to flower and fruit continuously until killed by frost. None was noted to emit an odor. British Guiana cultivars more nearly approached those of C. occidentalis in behavior, especially in the suffrutescent habit. Owing to the predations of an unidentified coleopterous insect, very little seed was set. However, seed broadcast on dry ground at the end of the rainy season did germinate when the rains resumed.
In both Texas and British Guiana, corolla wilt in C. sophera occurred some hours later, in no case before noon. The purple stem color used to distinguish C. sophera var. purpurea (Roxb.) Benth. is striking in living plants (all from Australia), but the character is lost in drying. In addition, the character is hardly absolute; garden plants of both species and field specimens of C. occidentalis all exhibited in varying degrees violescent stems, inflorescences, and fruit.
Data presented in Table 2 were derived from natural populations of C. occidentalis studied in the field in British Guiana, Mexico, and Texas, and from cultivated plants of both species grown in Austin, Texas, and Georgetown, British Guiana. In addition, the following relevant exsiccatae were examined together with the cited Suriname collection:
C. occidentalis. VENEZUELA. Vicinity of Cristobal Colon, Broadway 16; Lower Orinoco, Rushy & Squires 13. BRITISH GUIANA. Georgetown, Cox & Hubbard 148; Pomeroon Distr., Moruka R., de la Cruz 1075; Georgetown, Forest Dept. of British Guiana 6407; Georgetown, Harrison 1527; Waranama, Harrison & Persaud 1106; Vreed-en-Hoop, Hitchcock 16717; Georgetown, Irwin 1408; Coast region, Jenman 4408. SURINAME, w/o locality, M. Berthoud-Coulon 149; w/o locality, Hostmann 400; Saramacca R., Maguire 28893a. FRENCH GUIANA. Vicinity of Cayenne, Broadway 160.
C. sophera. CURACAO, w/o locality, Boldingh 5306. VENEZUELA. Caracas, Pittier 9712; D. F., near Caracas, Pittier 11563; GUADELOUPE. Basseterre, Pere Duss 2622. MARTINIQUE. Sumentin, R. Lizard, Pere Duss 826, 1835. BARBADOS. Doss, Bovell 45. TRINIDAD. Port of Spain, Broadway 8814; w/o locality, Fendler 335; w/o locality, Schack s.n. Bentham’s14 discussion on the origin and relationships of C. sophera suggest that the complexities of morphological variability and wide distribution, especially as they have been further confused by unintentional introduction in many places, could well be analyzed by computer techniques.